Deposit subscription

Content

• Summary
• Project description
• Data and methods
• Project structure

---

Summary

It was studied factors, affecting customers to subscribe for a term deposit after the direct marketing campaign. It identified the most significant factors of positive campaign outcomes. The customer may get a forecast of contact while marketing the campaign by a given data using the application application on a web server. The sample page shown below.

Microservice application page

Project description

A bank initiated an one-year marketing campaign among its clients to deposit money. After it was finished, the management decided to identfy success factors of the campaign using machine learning algorythm to increase its efficiency and decrease costs.

Data and methods

⬆️ to content

It might be suggested to contact customers either in December or in summer, because the probability of success depends on month:

It was studied several models such as linear regression, random forest, decision tree, gradient boosting and stacking

The model shows good predictive capability: it classifies correctly 78% of entries.

ROC curve

Most significant featurs are: poutcome_success, balance, age, pdays, contact by cellular phone, and others. It implies that elder people as well as young have propensity to subscribe for term depostit in case the previous contact was successful. The probability of success increases in late summer. Most people in these groups have secondary education and married. Meanwhile single people with tertiary education have higher rate to subscribe than that of other social group.

Project structure

display project structure

gesture_classification
├── .gitignore  
├── config
│   └── config.json           # configuration setings
├── data                      # data archive
│   └── campaign.zip
├── figures
│   ├── fig_1.png
.....
│   └── fig_streamlit.PNG
├── main.py
├── models                    # models and weights
│   ├── models_collection.py
│   ├── model_rf_opt.pkl
│   └── __ init __.py
├── notebooks                 # notebooks
│   └── Project_en.ipynb
├── project tree.ipynb
├── README.md                 # readme in English
└── utils                     # functions, variables, and data loaders
    ├── application.py
    ├── functions.py
    ├── reader_config.py
    ├── __ init __.py
    └── __pycache__

⬆️ to content